Date of Award

Summer 8-2020

Document Type


Degree Name

Doctor of Philosophy (PhD)


Computer Science

Committee Director

Stephan Olariu

Committee Member

Ravi Mukkamala

Committee Member

Michele C. Weigle

Committee Member

Dimitrie C. Popescu


Motivated by the success of the conventional cloud computing, Vehicular Clouds were introduced as a group of vehicles whose corporate computing, sensing, communication, and physical resources can be coordinated and dynamically allocated to authorized users. One of the attributes that set Vehicular Clouds apart from conventional clouds is resource volatility. As vehicles enter and leave the cloud, new computing resources become available while others depart, creating a volatile environment where the task of reasoning about fundamental performance metrics becomes very challenging. The goal of this thesis is to design an architecture and model for a dynamic Vehicular Cloud built on top of moving vehicles on highways. We present our envisioned architecture for dynamic Vehicular Cloud, consisting of vehicles moving on the highways and multiple communication stations installed along the highway, and investigate the feasibility of such systems. The dynamic Vehicular Cloud is based on two-way communications between vehicles and the stations. We provide a communication protocol for vehicle-to-infrastructure communications enabling a dynamic Vehicular Cloud. We explain the structure of the proposed protocol in detail and then provide analytical predictions and simulation results to investigate the accuracy of our design and predictions. Just as in conventional clouds, job completion time ranks high among the fundamental quantitative performance figures of merit. In general, predicting job completion time requires full knowledge of the probability distributions of the intervening random variables. More often than not, however, the data center manager does not know these distribution functions. Instead, using accumulated empirical data, she may be able to estimate the first moments of these random variables. Yet, getting a handle on the expected job completion time is a very important problem that must be addressed. With this in mind, another contribution of this thesis is to offer easy-to-compute approximations of job completion time in a dynamic Vehicular Cloud involving vehicles on a highway. We assume estimates of the first moment of the time it takes the job to execute without any overhead attributable to the working of the Vehicular Cloud. A comprehensive set of simulations have shown that our approximations are very accurate. As mentioned, a major difference between the conventional cloud and the Vehicular Cloud is the availability of the computational nodes. The vehicles, which are the Vehicular Cloud's computational resources, arrive and depart at random times, and as a result, this characteristic may cause failure in executing jobs and interruptions in the ongoing services. To handle these interruptions, once a vehicle is ready to leave the Vehicular Cloud, if the vehicle is running a job, the job and all intermediate data stored by the departing vehicle must be migrated to an available vehicle in the Vehicular Cloud.